Homework	Description
HW1	Train decision trees and random forests on the madelon and satimage datasets, and plot training and test misclassification errors.
HW2	Perform regressions on the abalone dataset to predict age from physical measurements, evaluate models using metrics like MSE and R^2.
HW3	Implement MAP learning for logistic regression using gradient descent and analyze the Gisette, madelon, and dexter datasets.
HW4	Train TISP classifiers on the Gisette, madelon, and dexter datasets using variable selection methods and plot various evaluation metrics.
HW5	Implement the FSA variable selection method for linear models and binary classification using the Lorenz loss on the Gisette, dexter, and madelon datasets.
HW6	Train a Logitboost classifier using univariate linear regressors as weak learners on the arcene, dexter, and Gisette datasets.
HW7	Train regression Neural Networks to predict pixel values from their coordinates in an image of a bird using different numbers of hidden layers and neurons.
HW8	Experiment with k-means and EM clustering on generated datasets to analyze clustering performance using metrics like accuracy and the Adjusted Rand Index.
HW9	Implement the Viterbi and Forward-Backward algorithms for a hidden Markov model to predict the most probable sequence of hidden states given an observed sequence. Also implement the Baum-Welch algorithm to learn the model parameters iteratively.
HW10	Perform spectral clustering on pixel data from an image, and display clustering results as well as mean color images for clusters.
HW11	Perform PCA on horse images and a bird image, discard the largest singular values, plot singular values, project images to PCs, and calculate distances to the PCA plane. Also, perform binary reconstruction of images using PCA.
HW12	Train and evaluate SVMs with polynomial and RBF kernels on various datasets (hill valley, sat, madelon, Gisette) to analyze misclassification errors across different hyperparameters.